Influence of Surface Protonation on the Sensitization Efficiency of

The influence of electrolyte pH on the sensitization efficiency of porphyrin-derivatized TiO2 was studied by photoelectrochemical and transient absorp...
2 downloads 0 Views 252KB Size
Download PDF
11680

J. Phys. Chem. B 2004, 108, 11680-11688

Influence of Surface Protonation on the Sensitization Efficiency of Porphyrin-Derivatized TiO2 David F. Watson, Andras Marton, Arnold M. Stux, and Gerald J. Meyer* Departments of Chemistry and Materials Science and Engineering, Johns Hopkins UniVersity, 3400 North Charles Street, Baltimore, Maryland 21218 ReceiVed: April 26, 2004

The influence of electrolyte pH on the sensitization efficiency of porphyrin-derivatized TiO2 was studied by photoelectrochemical and transient absorption measurements. The porphyrins 5-(4-carboxyphenyl)-10,15,20trimesitylporphinatozinc(II) (1), 5-(4-carboxyphenyl)-10,15,20-trimesitylporphine (2), 5-(4-carboxyphenyl)10,15,20-trimesitylporphinatoplatinum(II) (3), and 5-(4-dihydroxyphosphorylphenyl)-10,15,20-trimesitylporphinatozinc(II) (4) were anchored to low-surface-area and nanocrystalline TiO2 films. The TiO2 conduction band edge potential (ECB) shifted 59 ( 2 mV/pH, from -0.43 V vs Ag/AgCl(aq) at pH 12 to +0.16 V vs Ag/AgCl(aq) at pH 2. Excited-state potentials (E1/2(P+/*)) of 1-4 ranged from -1.58 to -0.91 V vs Ag/AgCl(aq), well negative of ECB. Despite the thermodynamic favorability of electron injection, a 10-fold increase in sensitized photocurrent was measured for 1-4 upon acidification of the electrolyte from pH 10 to 4. Transient absorption data revealed that sensitization of nanocrystalline TiO2 by 1-4 depended on pH in an identical manner. A mechanism is proposed wherein protonation of a surface site is required for charge compensation of injected electrons. Thus, the magnitude of sensitized photocurrent is determined by surface protonation-deprotonation equilibria.

Introduction Dye-sensitized photoelectrochemical cells have been the subject of considerable interest over the past three decades.1-5 The accepted mechanism for dye-sensitization involves the excitation of a sensitizer followed by interfacial electron transfer, or electron injection, into a semiconductor acceptor state.1,2,6-8 In regenerative dye-sensitized photoelectrochemical cells, the oxidized sensitizer is reduced by a donor present in the electrolyte, which is in turn reduced at the counter electrode. Thus, photocurrent and photovoltage are generated under subbandgap illumination, while no net chemistry occurs. The development of efficient dye-sensitized solar cells using singlecrystal electrodes was precluded by poor light harvesting efficiencies and low photocurrent densities. To improve the light harvesting efficiency, Gra¨tzel and co-workers9-12 developed high surface area, nanocrystalline dye-sensitized metal oxide films. In 1991, O′Regan and Gra¨tzel13 reported a global power conversion efficiency of ∼7% for a dye-sensitized nanocrystalline TiO2 electrode under simulated AM 1.5 solar illumination. The efficiency has since been optimized to 10.96%.14,15 The cation concentration at the semiconductor-electrolyte interface profoundly influences the efficiency of dye-sensitized photoelectrochemical cells.16 Cations have been found to affect many parameters, including the energetics of the semiconductor and sensitizer,17-22 the efficiency and dynamics of electron injection,23-26 and the kinetics of charge transport.27,28 Understanding and controlling cation effects is important with regard to maximizing the efficiency of dye-sensitized solar cells. In the early dye-sensitization literature, several reports described pH-dependent sensitized photocurrents at singlecrystal metal oxide electrodes.29-31 The effect was attributed to the well-known Nernstian dependence of the conduction band edge potential on the electrolyte proton concentration,17-19 and the resulting pH dependence of the driving force for electron

CHART 1: Structures of the Porphyrin Sensitizers, 1: M ) Zn(II), X ) CO2H; 2: M ) 2H, X ) CO2H; 3: M ) Pt(II), X ) CO2H; 4: M ) Zn(II), X ) PO3H2

injection. We recently reported pH-dependent sensitized photocurrents for porphyrin-modified, low-surface-area TiO2 films.32 This surprising result could not be explained by pH-induced shifts of the TiO2 conduction band, because electron injection was thermodynamically favorable over the entire pH range through which the photocurrent onset occurred. Instead, the result was attributed to the influence of surface protonationdeprotonation equilibria on the collection efficiency of injected electrons in low-surface-area TiO2 films. In this manuscript, we report on the continuation of our recent studies. We have characterized the sensitization of nanocrystalline and low-surface-area TiO2 films with porphyrins 1-4 (Chart 1) using photoelectrochemical and spectroscopic techniques. The interfacial proton concentration was found to exert similar effects on the sensitization efficiency of nanocrystalline TiO2 and low-surface-area TiO2. Our findings have important implications for optimizing the efficiency of nanocrystalline dyesensitized photoelectrochemical cells.

10.1021/jp048182x CCC: $27.50 © 2004 American Chemical Society Published on Web 07/01/2004

J. Phys. Chem. B, Vol. 108, No. 31, 2004 11681

Influence of Surface Protonation on TiO2 Experimental Section Materials. Porphyrin Synthesis. Porphyrins 1, 2, and 4 were obtained from Prof. J. S. Lindsey’s laboratory.33,34 Porphyrin 3 was synthesized by Pt2+ insertion into the free base porphyrin (2), following the method of Mink et al.35 Semiconductor Film Preparation. Single-crystal rutile TiO2 (100) was obtained from Princeton Scientific. Crystals were reduced by heating at 600 °C for 1 h in a 10/90 H2/N2 atmosphere. Low-surface-area TiO2 films were prepared by adaptation of the method of Ting et al.36,37 Elemental Ti was sputter-deposited onto conductive FTO-coated glass slides (8-20 Ω/square, Libby Owens Ford) at room temperature under a 10 mTorr Ar atmosphere at 50 W. Optimal film thickness was achieved with 2-6 min sputtering time. Ti films were oxidized to TiO2 by heating at 450 °C in air for 30-60 min. Nanocrystalline TiO2 and ZrO2 films on glass or FTO substrates were prepared as previously described.38,39 Surface Attachment Reactions. TiO2 films were exposed to solutions of 1-4 in toluene or THF for 12-15 h. Upon removal from porphyrin solutions, derivatized slides were rinsed thoroughly with the derivatizing solvent. Surface coverages were determined from the Soret and Q-band absorbances of derivatized films. Scanning Electron Microscopy. SEM images were obtained with a JEOL scanning electron microscope. Secondary electron images were acquired at 2-10 kV. Electrochemistry. Impedance Spectroscopy. Impedance and capacitance of low-surface-area TiO2/FTO samples were measured using a Stanford Research SR530 lock-in amplifier (LIA) and a PAR Model 173 potentiostat/galvanostat. All measurements were performed in a three-electrode configuration using a low-surface-area TiO2/FTO working electrode, Pt mesh counter electrode, and Ag/AgCl(aq) reference electrode. The electrolyte was aqueous 0.1 M Na2B4O7, the pH of which was adjusted with HCl and NaOH. The geometry of the electrochemical cell was the same throughout the experiments, enforcing a minimal distance between electrodes and a 0.28 cm2 working electrode area. Using a LabView program, the LIA and potentiostat were programmed to obtain frequency vs impedance data and potential vs capacitance data. The LIA was used to generate both the reference signal and the modulating signal (3 mV peak to peak) of known frequency. The modulating signal was added onto the potential generated by the potentiostat. The current to voltage output of the potentiostat was connected to the voltage input of the LIA. The LIA was then used to measure the real and imaginary part of the current in the electrochemical cell as a function of either frequency or potential. The real and imaginary impedance (Z) and capacitance (C) were calculated by the program. Complex plane plots of the impedance (|Im(Z)| vs Re(Z)), Bode plots (log(|Z|) vs log(ω)), and C vs potential plots were also produced by the LabView program. Photocurrent Onset Potentials. Band gap photocurrents were measured in a 3-electrode, single compartment cell with a lowsurface-area TiO2/FTO working electrode, Pt-coated FTO counter electrode, and Ag/AgCl(aq) reference electrode. The electrolyte was aqueous 0.1 M Na2B4O7, the pH of which was adjusted with HCl and NaOH. Linear sweep voltammograms (V ) 10 mV/s) were measured in the dark and under chopped (320 Hz) illumination by the full spectral output of a Xe lamp. Photocurrents were measured by lock-in technique with a Stanford Research Systems SR530 lock-in amplifier. Cyclic Voltammetry. Cyclic voltammetry was performed with a Bioanalytical Systems CV-50 potentiostat. A 3-electrode,

single-compartment cell was used, with a porphyrin-derivatized TiO2 working electrode, Pt mesh counter electrode, and Ag/ AgCl(aq) reference electrode. The electrolyte was 0.1 M tetrabutylammonium perchlorate in CH3CN or 0.1 M Na2B4O7 in water. The pH of the aqueous electrolyte was adjusted with HCl and NaOH. Sensitized Photocurrent Measurement. Photocurrents were measured using one of two setups. (1) A 3-electrode sandwich cell was employed, with a porphyrin-modified TiO2/FTO working electrode, Pt-coated FTO counter electrode, and Ag pseudoreference electrode. The electrolyte was 0.1 M aqueous Na2B4O7 with 0.05 M hydroquinone as an electron donor. The electrolyte pH was adjusted with HCl and NaOH. The working electrode was illuminated through the backside using a Schoeffel CPS 255HR 450 W Xe lamp coupled to a Kratos GM252 monochromator. Current was measured with a PAR 173 potentiostat output to a Houston Instruments 2000 X-Y recorder. Photocurrent action spectra were measured at constant positive bias at potentials corresponding to saturation sensitized photocurrent, as determined from linear sweep voltammetry measurements under chopped (∼1 Hz) Soret band illumination. (In linear sweep experiments, potential was controlled with a PAR 175 universal programmer.) (2) A 2-electrode sandwich cell was employed, with a porphyrin-modified TiO2/FTO working electrode and Pt-coated FTO counter electrode. The working electrode was illuminated through the backside using a 150 W Spectra-Physics Xe lamp coupled to an Oriel Cornerstone 1/4m monochromator. Current was measured by a Keithley 617 electrometer. Spectroscopy. Static Absorption and Emission. UV/vis absorption spectra were obtained with a Hewlett-Packard 8453 diode array spectrophotometer. Emission spectra were obtained with a Spex Fluorolog fluorimeter. Porphyrin-modified films were placed diagonally in a 1 cm cuvette, and front-face emission was measured. Transient Absorption Spectroscopy. Transient absorption data were acquired as described previously.24 The frequency-tripled, 355 nm output of a Q-switched Nd:YAG laser (Continuum Surelite) was Raman-shifted with H2(g), yielding pulsed 417 nm irradiation (∼8 ns, 5-10 mJ/cm2, 1 Hz). Sensitizer-derivatized semiconductor films were placed in a 1-cm quartz cuvette at a 45° angle to the pump beam. The cuvette was stoppered, and the surrounding aqueous electrolyte solution was purged with Ar or N2. The pump beam was expanded to an ∼1.5-cm diameter to ensure uniform illumination of the sensitizerderivatized films. The probe source was a pulsed 150 W Xe lamp (Applied Photophysics) focused on the semiconductor film. Transmitted light was output through a Spex 1702-04 monochromator and detected with a Hamamatsu R928 PMT. The pump and probe beams were orthogonal. The PMT signal was digitized by a LeCroy 9450 oscilloscope. Results TiO2 Morphology. Two types of TiO2 film were prepared: “low-surface-area” films from thermal oxidation of sputterdeposited titanium, and nanocrystalline films from sol-gel deposition. Scanning electron micrographs of low-surface-area TiO2 films (Figures 1a-c) reveal that the morphology of the films depended on the sputter deposition time scale. At short deposition times (e4 min), the TiO2 films consisted of relatively large particles with dimensions on the order of 100 nm. Some additional surface roughness on the individual TiO2 particles is apparent. The morphology of these films corresponded closely to that of their FTO substrates. Increasing the sputter deposition

11682 J. Phys. Chem. B, Vol. 108, No. 31, 2004

Watson et al. from Mott-Schottky analysis of impedance spectroscopy data (Figure 3a) and from the onset potentials of band gap photocurrent.19,43,44 Flat-band potentials determined from both methods were in close agreement (Figure 3b). Therefore, significant band-edge unpinning does not occur under illumination. MottSchottky plots (1/Cp2 vs V) were obtained at 10 and 100 Hz, which fall within the capacitive regime of the Bode plot at each electrolyte pH. The conduction band edge potential (ECB) was estimated from EFB:

ECB ) EFB - kBT ln

Figure 1. Scanning electron micrographs of TiO2 films on FTO. Images a-c are low-surface-area TiO2 films prepared by oxidation of sputter-deposited titanium films at deposition times of 4 min (a), 7.5 min (b), and 30 min (c). Image d is a nanocrystalline film from the sol-gel preparation.

time resulted in increased film thickness and increased surface area through the formation of smaller TiO2 particles. The film prepared by oxidation of 30-minute sputtered titanium consisted of aggregates of TiO2 with features on an approximately 10nm length scale. The nanocrystalline TiO2 film also consisted of aggregates of particles with roughly 10-nm dimensions (Figure 1d). Sensitizer Adsorption. Porphyrins 1-4 were adsorbed to TiO2 surfaces from solution. Equilibrium binding was welldescribed by the Langmuir adsorption isotherm model (Figure 2).40 Porphyrins were adsorbed to low-surface-area TiO2 films (2-6 min sputter deposition time) with surface adduct formation constants (Kad) of 105 M-1 and saturation surface coverages (Γmax) of 4.0 × 10-10 mol/cm2. For nanocrystalline TiO2 films, Kad ) 104 M-1 and Γmax ) 9.7 × 10-8 mol/cm2. Attempts to adsorb the tetramesityl analogue of 1 to TiO2 yielded no measurable surface coverage, suggesting that 1-4 are anchored to the surface through carboxylic acid or phosphonic acid functional groups. When low-surface-area TiO2 films derivatized with 1, 2, or 3 were immersed in organic solvents in which the porphyrins were soluble, rapid and nearly complete desorption occurred. Low-surface-area TiO2 derivatized with 4, which contains the phosphonic acid surfaceattachment group, underwent little or no desorption in organic solvents. For all four porphyrins, very little desorption occurred from nanocrystalline TiO2 films immersed in organic solvents. Porphyrins 1-4 remained attached to low-surface-area TiO2 when the films were immersed in water. The stable surface attachment is attributed to the insolubility of the porphyrins in water. The absorption spectra of porphyrin-derivatized TiO2 films immersed in aqueous solution were independent of pH, indicating that the porphyrins were stable toward diacid formation.41,42 The absorption and emission spectra of surface-attached porphyrins were minimally perturbed from their solution spectra. The absorption and emission spectra of 1 and 4, which differ in the nature of the surface-attachment group, were superimposable. These observations suggest that electronic coupling is relatively weak through the phenylene spacer connecting the porphyrin ring and the surface-attachment group. Semiconductor Energetics. The flat-band potentials (EFB) of underivatized, low-surface-area TiO2 films were determined

() nD NC

(1)

where nD is the donor density and NC is the density of states at the conduction band edge.30,45,46 Donor densities of 1018 cm-3 were calculated from Mott-Schottky slopes. Assuming that NC ∼ 1020 cm-3, ECB is no more than 120 mV negative of EFB. EFB values derived from Mott-Schottky plots shifted linearly with pH, with the expected Nernstian slope of 59 ( 2 mV/pH. From pH 2 to 12, EFB varied from +0.16 to -0.43 V vs Ag/ AgCl(aq). Impedance experiments were performed on 4-modified low-surface-area TiO2 at pH 10 and pH 12. (At more acidic pH’s, porphyrin oxidation current prevented impedance measurement.) EFB values of 4-modified TiO2 films were within 100 mV of those of unmodified TiO2. Sensitizer Energetics. Porphyrin ground-state reduction potentials (E1/2(P+/0)) were determined by cyclic voltammetry. Porphyrins 1 and 4, adsorbed to low-surface-area TiO2, underwent reversible one-electron ring oxidation. At pH 1, the E1/2(P+/0) values of 1 and 4 were +0.71 and +0.68 V vs Ag/ AgCl. The E1/2(P+/0) values of 1 and 4 were weakly pHdependent, shifting 15 ( 6 mV/pH and 10 ( 1 mV/pH, respectively. These shifts are consistent with previously reported measurements on phthalocyanine dyes attached to nanocrystalline TiO2.47 Reversible one-electron oxidation was observed for 1 adsorbed to nanocrystalline TiO2, with identical E1/2(P+/0) values as on low-surface-area TiO2. The E1/2(P+/0) values of 2 and 3 were more positive than the E1/2(P+/0) values of 1 and 4, and were outside the aqueous solvent window. In acetonitrile electrolyte, 2 adsorbed to low-surface-area TiO2 underwent quasireversible one-electron oxidation with E1/2(P+/0) ) +1.09 V vs Ag/AgCl. TiO2-adsorbed 3 underwent irreversible oneelectron oxidation with Eox ) +1.16 V vs Ag/AgCl. An E1/2(P+/0) value of +1.08 V vs Ag/AgCl was estimated by assuming a similar peak-to-peak separation as observed for 3. The measured E1/2(P+/0) value of TiO2-adsorbed 3 is roughly 100 mV positive of reported values for Pt(II) tetraphenylporphyrin in solution.35,48 The excited-state reduction potentials (E1/2(P+/*)) of surfacebound porphyrins were calculated using the standard RehmWeller equation,49 where E00 values were estimated from the emission spectra of porphyrins bound to nanocrystalline ZrO2 films. The emission spectra of porphyrin-modified ZrO2 films immersed in aqueous solution were pH-independent. The E1/2(P+/*) values of 1-4 span a range of 670 mV. The potential difference between E1/2(P+/*) and ECB determines the thermodynamic favorability of electron injection. For all four porphyrins, E1/2(P+/*) is more than 450 mV more negative of ECB, even at the most basic pH’s where ECB is most negative. Sensitizer energetics are summarized in Table 1. Photoelectrochemistry. Anodic photocurrents were measured for porphyrin-modified, low-surface-area TiO2 electrodes. Photocurrent action spectra (Figure 4) agreed with absorptance spectra. The magnitude of sensitized photocurrent increased by about

Influence of Surface Protonation on TiO2

J. Phys. Chem. B, Vol. 108, No. 31, 2004 11683

Figure 2. Equilibrium binding data for 1 on low-surface-area TiO2 (a) and nanocrystalline TiO2 (b). The insets show plots of [1]/Γ vs [1] and fits to the Langmuir adsorption isotherm. For low-surface-area TiO2, Kad ) 105 M-1 and Γmax ) 4.0 × 10-10 mol/cm2. For nanocrystalline TiO2, Kad ) 104 M-1 and Γmax ) 9.7 × 10-8 mol/cm2. Superimposed on the plots of Γ vs concentration are fits to the data based on the calculated values of Kad and Γmax.

Figure 3. (a) Mott-Schottky plots for an underivatized low-surface-area TiO2 film with frequency of 100 Hz and peak to peak voltage oscillation of 3 mV. Electrolyte pH ) 1.8 (0), 3.8 (O), 4.9 (4), 6.0 (3), 7.0 ()), 7.8 (left-facing triangle), 8.9 (right-facing triangle), 9.9 (9), and 10.7 (1). Superimposed are linear fits of the data. (b) TiO2 flat-band potential (EFB) vs electrolyte pH for an underivatized low-surface-area TiO2 film. EFB was estimated from the onset potentials of band gap photocurrent (0, slope ) 68 ( 3 mV/pH) and from the x-intercepts of Mott-Schottky plots (2, slope ) 59 ( 2 mV/pH).

TABLE 1: Energetics of Surface-Bound Porphyrins porphyrin

E1/2(P+/0) (V)a,b

E00 (eV)

E1/2(P+/*) (V)a,b

|E1/2(P+/*) - ECB|, pH 12 (eV)c,d

1 2 3 4

+0.54 +1.09 +1.08 +0.57

2.11 2.10 1.99 2.15

-1.57 -1.01 -0.91 -1.58

1.14 0.58 0.48 1.15

a For 1 and 4, pH 12 data are reported. b Potentials are reported vs Ag/AgCl. c ECB of low-surface-area TiO2 at pH 12 ) -0.43 V vs Ag/ AgCl. d |E1/2(P+/*) - ECB| indicates the thermodynamic favorability of electron injection, and is lowest at basic pH’s where ECB is most negative.

10-fold at pH < 5, independent of which porphyrin was used. Plots of photocurrent vs pH for 1-4 were nearly identical, despite the 670 mV range in E1/2(P+/*) values (Figure 4a, inset). The data were independent of whether surface-attachment was through the carboxylic acid or phosphonic acid surface attachment group. No changes were observed in the absorption spectra or photoaction spectra of porphyrin-derivatized TiO2 electrodes following photoelectrochemical measurements. The observed proton dependencies were reversible and independent of hydroquinone donor concentration. The decrease in photocurrent at

basic pH’s cannot be attributed to decreased hydroquinone oxidation kinetics, because hydroquinone oxidation is more favorable at high pH.50 Control experiments using porphyrinmodified single-crystal TiO2 or using iodide as the electron donor revealed the same pH dependence. Attempts to measure the pH dependence of sensitized photocurrent for porphyrinmodified nanocrystalline TiO2 films were complicated by the observation of cathodic sensitized photocurrents at acidic pH. Short-circuit photocurrents (isc) were linear with respect to incident photon flux (PF), and open-circuit photovoltages (Voc) were linear with respect to log(PF). Voc increased 65-100 mV per decade increase of PF, corresponding to an “ideality factor” of 1.1 to 1.7.51 With hydroquinone as the electron donor species, Voc was independent of pH in acidic electrolytes (pH < 5) where maximum sensitized photocurrents were measured. Transient Absorption Spectroscopy. Transient absorption data were obtained for 1-modified nanocrystalline TiO2 and ZrO2 films immersed in 0.1 M Na2B4O7 at pH 2.1, with excitation into the Soret band at 417 nm (Figure 5). The surface coverage (Γ ) of 1 was 3 × 10-9 mol/cm2 on both TiO2 and ZrO2. Transient absorbance difference spectra decayed with wavelength-independent kinetics. The spectrum of 1 on ZrO2

11684 J. Phys. Chem. B, Vol. 108, No. 31, 2004

Watson et al.

Figure 4. (a) Photocurrent action spectra of 2-derivatized low-surface-area TiO2 film at electrolyte pH’s of 2.0 (9), 3.0 (b), 4.4 (2), 5.8 (1), 7.2 ((), 8.7 (left-facing triangle), 10.2 (right-facing triangle), and 12.5 (0). (b) Photocurrent action spectra of 1-derivatized low-surface-area TiO2 film at electrolyte pH’s of 2.0 (9), 3.5 (b), 5.0 (2), 6.5 (1), 8.0 ((), 9.5 (left-facing triangle), 11.0 (right-facing triangle), and 12.5 (0). IPCE ) incident photon-to-current efficiency. Superimposed on the photoaction spectra are absorptance spectra of the films (-). Inset to Figure 4a: Plots of Soret band photocurrent vs electrolyte pH for 1(9), 2(b), 3(2), and 4(1).

Figure 5. Transient absorbance difference (∆A) spectra of 1-modified nanocrystalline TiO2 (9) and 1-modified nanocrystalline ZrO2 (0) films, 10 ns after 417 nm excitation. Γ1 ) 3 × 10-9 mol/cm2 on both films. The films were immersed in 0.1 M Na2B4O7 at pH 2.1. Superimposed on the spectrum of 1-modified TiO2 is a simulated spectrum (---) based on the sum of contributions from the 3(π,π*) excited state and the oxidized porphyrin.

consists of a broad absorption band overlapping the bleach of the ground-state Soret band, and a weaker absorption overlapping the Q(1,0) bleach. The spectrum is assigned to the 3(π,π*) excited-state, herein referred to as 1*. The 1(π,π*) excited-state lifetimes of Zn(II) tetraarylporphyrins are 108 s-1 and kcr > 108 s-1. Lian and co-workers65-68 have recently reported rapid nonexponential charge recombination for dye-sensitized nanocrystalline TiO2 with g70% of recombination occurring in less than 1 ns. Similarly, Willig et al.69 reported nonexponential recombination that was completed in less than 1.5 ns. Our proposed mechanism is also consistent with Lyon and Hupp’s70,71 observation that electrochemical reduction of nanocrystalline TiO2 films is accompanied by reversible proton uptake from aqueous solution. They have attributed the proton uptake to adsorption or intercalation driven by charge compensation of conduction band electron density. We speculate that proton intercalation may be coupled to electron transport at acidic pH’s. However, our photoelectrochemical and spectroscopic measurements do not enable us to directly probe the concentration of protons within the TiO2 film. The dependence of charge collection efficiency on electrolyte pH affects the energetic requirements for maximizing the global efficiency of dye-sensitized photoelectrochemical cells. The global efficiency is proportional to the short circuit photocurrent and the open circuit voltage. The short circuit photocurrent depends on φinj and ηel (eq 2). For both φinj and ηel to be simultaneously maximized, the entire excited-state distribution function (Wdon(E)) must lie at more negative potential than ECB at electrolyte proton concentrations sufficient for charge compensation of the injected electron. Open circuit voltage is maximized when Wdon(E) is only minimally more negative than ECB. It is often assumed that the pH dependence of ECB should enable some tunability of the open circuit voltage without affecting the short circuit photocurrent. However, the pH dependence of charge collection efficiency implies that this is not the case. The interplay between maximizing open circuit voltage and short circuit photocurrent imparts specific energetic requirements on the choices of semiconductor, sensitizer, and electrolyte composition. For example, the porphyrin sensitizers employed in this study are not ideal for maximizing the global efficiency of aqueous dye-sensitized TiO2 cells. At acidic pH’s, where maximum photocurrents were measured, E1/2(P+/*) values for 1-4 ranged from 1.1 to 1.6 eV negative of ECB. Electron injection is unnecessarily exergonic, resulting in unrecoverable losses of energy upon electron injection. However, raising the pH to minimize the free energy lost in the injection process fails, as this provides insufficient proton concentration for efficient charge collection. Conclusions We have reported pH-dependent photocurrents and transient absorption data for porphyrin-sensitized TiO2. Sensitized photocurrents were “turned off” at electrolyte pH’s above 10, even though electron injection was thermodynamically favored. Plots of photocurrent vs pH and ∆A0/∆A0,max vs pH were independent of the porphyrin used, despite the 670 mV range of E1/2(P+/*) values for 1-4. Identical pH-dependencies were observed for low-surface-area and nanocrystalline TiO2. The data indicate that the charge collection efficiency, and not the electron injection yield, is pH-dependent. We have proposed a mechanism in which trapping of the injected electron at TiIV-OH2+ surface sites leads to efficient charge collection, while trapping at TiIV-OH sites leads to efficient charge recombination. Thus, the charge collection efficiency is determined by surface

J. Phys. Chem. B, Vol. 108, No. 31, 2004 11687 protonation-deprotonation equilibria. Our findings have implications with regard to the design of efficient dye-sensitized photoelectrochemical cells. Judicious choices of semiconductor, sensitizer, and electrolyte are required to maximize the open circuit voltage while simultaneously optimizing the electron injection and charge collection efficiencies. Acknowledgment. The porphyrin complexes 1, 2, and 4 were provided by Mr. W. J. Youngblood, Dr. K. Muthukumaran, and Prof. J. S. Lindsey. We thank Profs. Lindsey, D. Holten, and D. Bocian for helpful discussions. This research was funded by the National Renewable Energy Laboratory. References and Notes (1) Gerischer, H. Photochem. Photobiol. 1972, 16, 243-260. (2) Memming, R. Photochem. Photobiol. 1972, 16, 325-333. (3) Hagfeldt, A.; Gra¨tzel, M. Chem. ReV. 1995, 95, 49-68. (4) Kalyanasundaram, K.; Gra¨tzel, M. Coord. Chem. ReV. 1998, 77, 347-414. (5) Qu, P.; Meyer, G. J. In Electron Transfer in Chemistry; Balzani, V., Ed.; John Wiley and Sons: New York, 2001; Vol. IV, pp 355-411. (6) Gerischer, H. Surf. Sci. 1969, 18, 97-122. (7) Gleria, M.; Memming, R. Z. Phys. Chem. (Munich) 1975, 98, 303316. (8) Memming, R. Surf. Sci. 1980, 101, 551-563. (9) Desilvestro, J.; Gra¨tzel, M.; Kavan, L.; Moser, J. J. Am. Chem. Soc. 1985, 107, 2988-2990. (10) Kalyanasundaram, K.; Vlachopoulos, N.; Krishnan, V.; Monnier, A.; Gra¨tzel, M. J. Phys. Chem. 1987, 91, 2342-2347. (11) Vlachopoulos, N.; Liska, P.; Augustynski, J.; Gra¨tzel, M. J. Am. Chem. Soc. 1988, 110, 1216-1220. (12) Liska, P.; Vlachopoulos, N.; Nazeeruddin, M. K.; Comte, P.; Gra¨tzel, M. J. Am. Chem. Soc. 1988, 110, 3686-3687. (13) O’Regan, B.; Gra¨tzel, M. Nature 1991, 353, 737-740. (14) Nazeeruddin, M. K.; Kay, A.; Rodicio, I.; Humphry-Baker, R.; Mu¨ller, E.; Liska, P.; Vlachopoulos, N.; Gra¨tzel, M. J. Am. Chem. Soc. 1993, 115, 6382-6390. (15) Gra¨tzel, M. In Future Generation PhotoVoltaic Technologies: AIP Conference Proceedings; McConnell, R. D., Ed.; NREL: Denver, CO, 1997. (16) Watson, D. F.; Meyer, G. J. Coord. Chem. ReV. In press. (17) Finklea, H. O. In Semiconductor Electrodes; Finklea, H. O., Ed.; Elsevier: New York, 1988; pp 43-145. (18) Watanabe, T.; Fujishima, A.; Honda, K. Bull. Chem. Soc. Jpn. 1976, 49, 355-358. (19) Bolts, J. M.; Wrighton, M. S. J. Phys. Chem. 1976, 80, 26412645. (20) Rothenberger, G.; Fitzmaurice, D.; Gra¨tzel, M. J. Phys. Chem. 1992, 96, 5983-5986. (21) Redmond, G.; Fitzmaurice, D. J. Phys. Chem. 1993, 97, 14261430. (22) Boschloo, G.; Fitzmaurice, D. J. Phys. Chem. B 1999, 103, 78607868. (23) Liu, Y.; Hagfeldt, A.; Xiao, X.-R.; Lindquist, S.-E. Sol. Energy Mater. 1998, 55, 267-281. (24) Kelly, C. A.; Farzad, F.; Thompson, D. W.; Stipkala, J. M.; Meyer, G. J. Langmuir 1999, 15, 7047-7054. (25) Tachibana, Y.; Haque, S. A.; Mercer, I. P.; Moser, J. E.; Klug, D. R.; Durrant, J. R. J. Phys. Chem. B 2001, 105, 7424-7431. (26) Asbury, J. B.; Anderson, N. A.; Hao, E.; Ai, X.; Lian, T. J. Phys. Chem. B 2003, 107, 7376-7386. (27) Kopidakis, N.; Schiff, E. A.; Park, N.-G.; van de Lagemaat, J.; Frank, A. J. J. Phys. Chem. B 2000, 104, 3930-3936. (28) Kambe, S.; Nakade, S.; Kitamura, T.; Wada, K.; Yanagida, S. J. Phys. Chem. B 2002, 106, 2967-2972. (29) Watanabe, T.; Fujishima, A.; Tatsuoki, O.; Honda, K.-i. Bull. Chem. Soc. Jpn. 1976, 49, 8-11. (30) Clark, W. D. K.; Sutin, N. J. Am. Chem. Soc. 1977, 99, 46764682. (31) Sonntag, L. P.; Spitler, M. T. J. Phys. Chem. 1985, 89, 14531457. (32) Watson, D. F.; Marton, A.; Stux, A. M.; Meyer, G. J. J. Phys. Chem. B 2003, 107, 10971-10973. (33) Lindsey, J. S.; Prathapan, S.; Johnson, T. E.; Wagner, R. W. Tetrahedron 1994, 50, 8941-8968. (34) Muthukumaran, K.; Loewe, R. S.; Ambroise, A.; Tamaru, S.-i.; Li, Q.; Mathur, G.; Bocian, D. F.; Misra, V.; Lindsey, J. S. J. Org. Chem. 2004, 69, 1444-1452. (35) Mink, L. M.; Neitzel, M. L.; Bellomy, L. M.; Falvo, R. E.; Boggess, R. K.; Trainum, B. T.; Yeaman, P. Polyhedron 1997, 16, 2809-2817.

11688 J. Phys. Chem. B, Vol. 108, No. 31, 2004 (36) Ting, C.-C.; Chen, S.-Y.; Liu, D.-M. J. Appl. Phys. 2000, 88, 46284633. (37) Ting, C.-C.; Chen, S.-Y.; Liu, D.-M. Thin Solid Films 2002, 402, 290-295. (38) Heimer, T. A.; D’Arcangelis, S. T.; Farzad, F.; Stipkala, J. M.; Meyer, G. J. Inorg. Chem. 1996, 35, 5319-5324. (39) Bonhoˆte, P.; Gogniat, E.; Tingry, S.; Barbe´, C.; Vlachopoulos, N.; Lenzmann, F.; Comte, P.; Gra¨tzel, M. J. Phys. Chem. B 1998, 102, 14981507. (40) Langmuir, I. J. Am. Chem. Soc. 1918, 40, 1361-1402. (41) Kynukshto, V. N.; Solovyov, K. N.; Egorova, G. D. Biospectroscopy 1998, 4, 121-133. (42) Chirvony, V. S.; van Hoek, A.; Galievsky, V. A.; Sazanovich, I. V.; Schaafsma, T. J.; Holten, D. J. Phys. Chem. B 2000, 104, 9909-9917. (43) Mott, N. F. Proc. R. Soc. London, Ser. A 1939, 171, 27. (44) Schottky, W. Z. Phys. 1942, 118, 539-592. (45) Gomes, W. P.; Cardon, F. Z. Phys. Chem. (Munich) 1973, 86, 330334. (46) Frank, S. N.; Bard, A. J. J. Am. Chem. Soc. 1975, 97, 7427-7433. (47) Zaban, A.; Ferrere, S.; Gregg, B. A. J. Phys. Chem. B 1998, 102, 452-460. (48) Tokel-Takvoryan, N. E.; Bard, A. J. Chem. Phys. Lett. 1974, 25, 235-238. (49) Rehm, D.; Weller, A. Isr. J. Chem. 1970, 8, 259-271. (50) Bailey, S. I.; Ritchie, I. M.; Hewgill, F. R. J. Chem. Soc., Perkin Trans. 2 1983, 645-653. (51) So¨dergren, S.; Hagfeldt, A.; Olsson, J.; Lindquist, S.-E. J. Phys. Chem. 1994, 98, 5552-5556. (52) Rodriguez, J.; Kirmaier, C.; Holten, D. J. Am. Chem. Soc. 1989, 111, 6500-6506. (53) Pekkarinen, L.; Linschitz, H. J. Am. Chem. Soc. 1960, 82, 24072411. (54) Felton, R. H. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978; Vol. 5, pp 53-125.

Watson et al. (55) Albery, W. J.; Bartlett, P. N. J. Electrochem. Soc. 1984, 131, 316325. (56) Finklea, H. O. In Semiconductor Electrodes; Finklea, H. O., Ed.; Elsevier: New York, 1988; pp 1-42. (57) Watanabe, T.; Fujishima, A.; Honda, K.-I. Chem. Lett. 1974, 897900. (58) Nozik, A. J. Annu. ReV. Phys. Chem. 1978, 29, 189-222. (59) Gerischer, H. In Physical Chemistry: An AdVanced Treatise; Eyring, H., Henderson, D., Jost, W., Eds.; Academic Press: New York, 1970; Vol. 9A, pp 463-542. (60) Stanienda, A.; Bieble, G. Z. Phys. Chem. (Munich) 1967, 52, 254275. (61) Giraudeau, A.; Callot, H. J.; Gross, M. Inorg. Chem. 1979, 18, 201-206. (62) Schindler, P. W.; Gamsja¨ger, H. Discuss. Faraday Soc. 1971, 52, 286-288. (63) Dobson, K. D.; Connor, P. A.; McQuillan, A. J. Langmuir 1997, 13, 2614-2616. (64) Connor, P. A.; Dobson, K. D.; McQuillan, A. J. Langmuir 1999, 15, 2402-2408. (65) Ghosh, H. N.; Asbury, J. B.; Weng, Y.; Lian, T. J. Phys. Chem. B 1998, 102, 10208-10215. (66) Weng, Y.; Wang, Y.-Q.; Asbury, J. B.; Ghosh, H. N.; Lian, T. J. Phys. Chem. B 2000, 104, 93-104. (67) Hao, E.; Anderson, N. A.; Asbury, J. B.; Lian, T. J. Phys. Chem. B 2002, 106, 10191-10198. (68) Wang, Y.; Hang, K.; Anderson, N. A.; Lian, T. J. Phys. Chem. B 2003, 107, 9434-9440. (69) Hannappel, T.; Burfeindt, B.; Storck, W.; Willig, F. J. Phys. Chem. B 1997, 101, 6799-6802. (70) Lyon, L. A.; Hupp, J. T. J. Phys. Chem. 1995, 99, 15718-15720. (71) Lyon, L. A.; Hupp, J. T. J. Phys. Chem. B 1999, 103, 46234628.